|
|
The banks assessed include the G-SIBs and the non-G-SIBs that were part of the sample of banks used to designate G-SIBs, as well as the additional sample of banks that complete the G-SIB template but do not rank among the 75 largest banks. The analysis has made use of data publicly disclosed by 104 banks with a Basel III leverage ratio exposure measure greater than EUR 200 billion euros in the 2013–17 period.
Several elements have been assessed. First, the numerators and denominators of the 12 G-SIB indicators were analysed separately. Then, the extent of regional differences in the behaviour of the GSIBs and non-G-SIBs were assessed.
The analysis of indicator numerators and denominators has revealed that both the group of GSIBs and that of non-G-SIBs are heterogeneous, with the implication that some central tendency measures within each group are highly influenced by a few banks. With regard to G-SIBs, the results suggest a general reduction in G-SIB scores over time for most indicators and for the final score. These developments are consistent with the objectives of the G-SIB framework. The observed reduction in scores becomes more evident when exchange rates and denominators are fixed. Exceptions to the observation of a reduction in indicator scores nevertheless occur with the indicators of the substitutability category, where a small increase has been observed or where the reduction is less evident. Finally, movements in exchange rates appear to have had a significant effect on the results.
The analysis reveals that non-G-SIBs behave differently from G-SIBs as a group. Non-G-SIBs have increased all of their G-SIB indicators as well as their final G-SIB score. This finding is once again consistent with the idea that the G-SIB framework may provide additional incentives to G-SIBs to reduce their G-SIB scores.
While the analysis finds that G-SIBs as a group have reduced their G-SIB scores, the regional analysis indicates heterogeneity across countries and regions. Euro area, GB and US G-SIBs have reduced their systemic importance, especially when exchange rate effects are removed. On the other hand, Chinese and Japanese G-SIBs have exhibited relatively positive growth rates for all indicators, and particularly high growth rates for the substitutability category indicators. At the same time, the regional analysis shows an increase in the relative importance of non-G-SIBs in all countries and regions.
While this paper makes a first attempt to analyse experience to date with the G-SIB framework, it is important to note that a number of potentially important factors have not been taken into account. Authors have not been able to assess the effect of business cycles in different countries, of market views on the effectiveness of the G-SIB framework, or the degree to which capital requirements may or may not be binding for banks in the sample. In addition, it was not possible at this juncture to disentangle the effects of other measures that have been put in place during the analysed period and which were aimed at GSIBs or non-G-SIBs with a Basel III leverage ratio exposure measure greater than EUR 200 billion. Finally, it was not possible at the current juncture to undertake an econometric exercise which could test hypotheses relating to G-SIB behaviour or to establish relationships between the G-SIB framework and observed behaviour. All of these factors are potentially relevant for assessing how far the G-SIB framework has achieved the intended objectives.